You are on page 1of 14

Communicative Tasks and the Language Curriculum

David Nunan
Over the last 25 years the communicative task has emerged as a
significant building block in the development of language curricula
and also as an element for motivating process-oriented second
language acquisition research. This paper reviews the influence of the
communicative task on curriculum development and summarizes the
research base for task-based language teaching. In the final part of
the paper, an agenda for future research is set out.
Over the last 25 years, the communicative task has evolved as an
important component within curriculum planning, implementation,
and evaluation. In task-based language teaching, syllabus content
and instructional processes are selected with reference to the
communicative tasks which learners will (either actually or
potentially) need to engage in outside the classroom and also with
reference to theoretical and empirical insights into those social and
psycholinguistic processes which facilitate language acquisition. This
approach to language teaching is characterized by the following
features;
1. An emphasis on learning to communicate through interaction in the
target language
2. The introduction of authentic texts into the learning situation
3. The provision of opportunities for learners to focus, not only on
language, but also on the learning process itself
4. An enhancement of the learner’s own personal experiences as
important contributing elements to classroom learning
5. An attempt to link classroom language learning with language
activation outside the classroom
Task-based language teaching has been an important addition to the
conceptual and empirical repertoire of the second and foreign
language teacher in the eighties, having influenced syllabus design,
materials development, and language teaching methodology. In this
paper, I shall review the development of task-based language
teaching (TBLT) Inthe first part of the paper, I shall provide an
account of the theoretical and empirical basis for TBLT. I shall then
discuss the influence of TBLT on curriculum development and
classroom practice. In the final part of the paper, I indicate the ways
in which I believe
that the research agenda should be extended in the nineties
THE CONCEPTUAL BASIS
Like many other innovations, task-based teaching entered the
language field from the educational mainstream. Studies of teachers
at work demonstrated that, while teacher education programs taught
trainees to plan, implement, and evaluate their programs according
to the “rational” model which begins with objectives and moves
through tasks to evaluation (Tyler, 1949), the reality was that once
they began practicing, teachers tended to focus on pedagogic tasks
(Shavelson & Stem, 1981). This insight from research into teachers
professional planning and decision-making processes enhanced the
status of task as a curriculum planning tool.
Task-based learning is also linked to mainstream education by its
close relationship with experiential learning. This relationship is
evident in the following description of experiential learning:
In experiential learning, immediate personal experience is seen as the
focal point for learning, giving “life, texture, and subjective personal
meaning to abstract concepts and at the same time providing a
concrete, publicly shared reference point for testing the implications
and validity of ideas created during the learning process,” as pointed
out by David KoIb (1984: 21). But experience also needs to be
processed consciously by reflecting on it. Learning is thus seen as a
cyclical process integrating immediate experience, reflection, abstract
conceptualization and action. (Kohonen, in press)
To date, definitions of tasks have been rather programmatic. Long
(1985a) suggests that a task is nothing more or less than the things
people do in everyday life. He cites as examples buying shoes,
making reservations, finding destinations, and writing cheques. The
Longman Dictionary of Applied Linguistics provides a more
pedagogically oriented characterization. Here, it is suggested that a
task is:
any activity or action which is carried out as the result of processing
or understanding language (i.e., as a response). For example, drawing
a map while listening to a tape, listening to an instruction and
performing a command, may be referred to as tasks. (Richards, Platt,
& Weber, 1985, p. 289)
The value of tasks, according to the authors, is that they provide a
purpose for the activity which goes beyond the practice of language
for its own sake.
A similar characterization is offered by Breen (1987) who suggests
that a task is:
any structured language learning endeavour which has a particular
objective, appropriate content, a specified working procedure, and a
range of outcomes for those who undertake the task. ‘Task’ is
therefore assumed to refer to a range of workplans which have the
overall purpose of facilitating language learning—from the simple and
brief exercise type, to more complex and lengthy activities such as
group problem- solving or simulations and decision making. (p. 23)
Elsewhere, I have suggested that tasks can be conceptualized in
terms of the curricular goals they are intended to serve, the input
data which forms the point of departure for the task, and the
activities or procedures which the learners undertake in the
completion of the task. Two important additional elements are the
roles for teachers and learners implicit in the task, and the settings
and conditions under which the task takes place (Nunan, 1989). Later
in this paper, I shall use these elements of goals, input data,
activities/procedures, roles, and settings as rubrics for synthesizing
the considerable amount of research activity which provides an
empirical basis for task-based language teaching and learning.
THE CURRICULAR BASIS
Before the development of communicative approaches to language
teaching, tasks and exercises were selected as a second order
activity, after the specification of the morphosyntactic, phonological,
and lexical elements to be taught. Traditionally, curriculum designers
and materials writers took as their point of departure the question,
What are the grammatical, phonological, and lexical items to be
taught? The specification of these items set the parameters for the
selection of classroom activities. In other words, selection of
classroom activities was driven by curriculum goals specified in
phonological, morphosyntactic, and lexical terms. (See, for example,
the analysis of
content selection and sequencing in a grammar- based syllabus
provided by McDonough, 1981, p.21.)
In a task-based curriculum, the decision-making process is quite
different. There are, in fact, two different routes which the curriculum
developer/materials writer can take ininitiating the design process.
The first of these is based on what I have called the rehearsal
rationale. Here the question initiating the design process is, What is it
that learners potentially or actually need to do with the target
language? The second is what I have called the psycholinguistic
rationale. Here the initiating question is, What are the
psycholinguistic mechanisms underlying second language acquisition,
and how can these
be activated in the classroom? The linguistic elements to be focused
on in the classroom are selected as a second order activity.
Ideally, task selection should occur with reference both to target task
rationale and psycholinguistic principles. The way that this might be
achieved is illustrated in the procedure set out in Figure 1, adapted
from a recently published task-based coursebook (Nunan &
Lockwood, 1991). The pedagogic task is selected with reference to
the realworld or target task of “giving information in a job interview.”
Learners are given a model of the target language behaviour, as well
as specific practice in manipulating key
language items. The actual pedagogic task, a simulation, is also
consistent with research on the facilitative effects of classroom
interaction (research on language acquisition is reviewed in the next
section).

FIGURE 1
Steps Involved in the Development of a Pedagogic Task
Procedure Example Rationale
1. Identify target task Giving personal To give learners the
information in a job opportunity to
interview develop
language skills
relevant to their real
world needs
2. Provide model Students listen to and To provide learners
extract key the
information opportunity to listen
from to and analyse ways
authentic/simulated in which native
interview speakers or users of
the target language
carry out the target
task
3. Identify enabling Manipulation drill to To provide learners
skill practice wh-questions with
with do-insertion explicit instruction
and
guided practice in
those
grammatical
elements
needed to perform
the target task
4. Devise pedagogic Interview simulation To provide learners
task using role cards the
opportunity to
mobilize
their emerging
language
skills through
rehearsal

The growing importance of the pedagogic task as a central element


within the curriculum has called into question the conventional
distinction between syllabus design and methodology. Traditionally,
syllabus design is concerned with the selection and grading of
content, while methodology is concerned with the selection and
sequencing of tasks, exercises, and related classroom activities.
Metaphorically speaking, syllabus design is concerned with the
destination, while methodology is concerned with the route. With the
development of task-based approaches to language learning and
teaching, this distinction has become difficult to sustain. Breen (1984)
has neatly captured this change of focus in the following way:
[TBLT would] prioritize the route itself; a focusing upon the means
towards the learning of a new language. Here the designer would give
priority to the changing process of learning and the potential of the
classroom—to the psychological and social resources applied to a new
language by learners in the classroom context. . . . a greater concern
with capacity for communication rather than repertoire of
communication, with the activity of learning a language viewed as
important as the language itself, and with a focus upon means rather
than predetermined objectives, all indicate priority of process over
content. (pp. 52-53)
Conceptually, then, task-based language teaching has been
influenced by developments in mainstream education as well as by
major conceptual shifts in our understanding of the nature of
language and language learning. It has also been enhanced by a
research agenda which has provided an empirical basis upon which
curriculum designers, materials writers, and classroom practitioners
can draw. The availability of empirical data on tasks has enhanced
the status of task-based language teaching at a time when the
various “methods” approaches to language teaching have come
under increasing criticism for
lacking an empirical basis. (See, for example, Long, 1990; Richards,
1990). In the next section, I shall provide a selective review of this
research.

THE EMPIRICAL BASIS


One of the strengths of task-based language teaching is that the
conceptual basis is supported by a strong empirical tradition. This
distinguishes it from most methods approaches to pedagogy, which
are relatively data-free. I have already suggested that tasks can be
conceptualized in terms of the key elements of goals, input data,
activities/procedures, roles, and settings. This conceptual scheme
provides a convenient means of synthesising the research on tasks.
Task goals enable the program planner and materials writer to
provide explicit links between the task and the broader curriculum it
is designed to serve. Without clearly articulated sets of goal
statements, there is a risk that task-based teaching programs will lack
coherence as Widdowson (1987), among others, has pointed out.
Goals are generally referenced against the sorts of things which
learners want to do with the language outside the classroom. Typical
goal statements include:
1. To develop the skills necessary to take part in academic study
2. To obtain sufficient oral and written skills to obtain a promotion
from unskilled worker to site supervisor
3. To communicate socially in the target language
4. To develop the survival skills necessary to obtain goods and
services
5. To be able to read the literature of the target culture
Despite its importance for coherent curriculum development,
compared to other areas, research on task goals is difficult to find in
the literature. One of the few available studies is that by Brindley
(1984) who investigated the needs analysis, goal and objective
setting practices of teachers of ESL to adults, and the reaction of
learners to these practices Based on an extensive series of
interviews, Brindley found that programs in which the goals were
explicit and reflected the communicative needs of the learners had
greater face validity than those in which the goals were either
unstated, inexplicit, or which did not reflect learners’ goals. While
there was no direct evidence that programs with explicit, relevant
goals resulted in more effective learning outcomes, it is not
unreasonable to expect that this would be the case, given what we
know about the relationship between affective/attitudinal factors and
learning outcomes.
Most tasks take as their point of departure input data of some sort.
Such data may be linguistic (that is, reading and listening texts of
various sorts) or nonlinguistic (for example, diagrams, photographs,
picture sequences). This area is considerably better researched than
that of goals. A key question underlying research on input tasks is,
What factors are implicated in the difficulty of aural and written texts?
In a large-scale investigation of the listening comprehension of
secondary students, Brown and Yule (1983) found that two factors
significantly affected the difficulty of listening texts. The first factor
related to the number of elements in the text and the ease and
difficulty of distinguishing between them. The second significant
factor was the text type. All other things being equal, descriptions
were easier than instructions, which were easier than stories.
Arguments or opinion-expressing texts containing abstract concepts
and relationships were the most difficult. Follow-up research cited in
Anderson and Lynch (1988) identified a number of other factors w
including the following:: :
1. The way the information is organized (narrative texts in which the
order of events in the texts mirrors the order in which the events
actually occurred in real life are easier to comprehend than narratives
in which the events are presented out of sequence)
2. The familiarity of the topic
3. The explicitness and sufficiency of the information
4. The type of referring expressions (for young children, pronominal
referents are more difficult to comprehend than full noun phrase
referents)
5. Text type
In the area of reading comprehension, Nunan (1984) found that
similar elements were implicated in the difficulty of school texts for
secondary level students. Nunan looked, among other things, at the
difficulty of different types of textual relationships as well as at the
effect of content familiarity. He found that logical relationships of the
type marked by conjunctions were more difficult than referential and
lexical relationships. He alsofound that content familiarity was more
significant than grammatical complexity in determining the difficulty
of reading texts.
The bulk of task-based research has focused on the activities or
procedures which learners carry out in relation to the input data. The
key question here has been, What tasks seem to be most helpful in
facilitating second language acquisition?
In the first of a series of investigations into learner-learner interaction,
Long (1981) found that two-way tasks (in which all students in a
group discussion had unique information to contribute) stimulated
significantly more modified interactions than one-way tasks (that is,
in which one member of the group possessed all the relevant
information). Similarly, Doughty and Pica (1986) found that required
information-exchange tasks generated significantly more modified
interaction than tasks in which the exchange of information
was optional. (Modified interactions are those instances in which
speakers modify their language in order to assure that they have
been correctly understood; they result from an indication of
noncomprehension usually on the part of a listener)
These investigations of modified interaction were theoretically
motivated by Krashen’s (1981, 1982) hypothesis that comprehensible
input was a necessary and sufficient condition for second language
acquisition—in other words, that acquisition would occur hen learners
understood messages in the target language. Long (1985b) advanced
the following argument in favor of tasks which promote
conversational adjustments or interactional modifications on the part
of the learners taking part in the task:
Step 1: Show that (a) linguistic/conversational adjustments promote
(b) comprehensible input.
Step 2: Show that (b) comprehensible input promotes (c) acquisition.
Step 3: Deduce that (a) linguistic/conversational adjustments promote
(c) acquisition.
Satisfactory evidence of the a-b-c relationships would allow the
linguistic environment to be posited as an indirect causal variable in
SLA. (The relationship would be indirect because of the intervening
“comprehension” variable.) (p. 378)
In the fast few years the comprehensible input hypothesis has been
criticised on
theoretical and empirical grounds. For example, Swain (1985)
demonstrated that immersion programs in Canada, in which learners
received huge amounts of comprehensible input did not lead to the
sort of native-like facility in the target language predicted by the
input hypothesis. She proposed that in addition to comprehensible
input, learners need opportunities that require that :heir own speech
be comprehensible because it is only through such opportunities that
learners are pushed to mobilize their emerging grammatical
competence. (Such mobilization is precisely what the tasks suggested
by Long, 1985b; Doughty & Pica, 1986; and others manage to
achieve. In other words, their research may be justified on grounds
other than that proposed by the comprehensible input hypothesis.)
More recently, attention has focused on the question of the types of
language and discourse patterns stimulated by different task types.
Berwick (1988, in press) investigated the different types of language
stimulated by transactional and interpersonal tasks. (A transaction
task is one in which communication occurs principally to bringabout
the exchange of goods and services, whereas an interpersonal task is
one in which communication occurs largely for social purposes.) He
found that the different functionalpurposes stimulated different
morphosyntactic realizations.
In a recent study, I investigated the different interactional patterns
stimulated by open and closed tasks. (An open task is one in which
there is no single correct answer, while a closed task is one in which
there is a single correct answer or a restricted number of correct
answers.) It was found that the different task types stimulated very
different interactional patterns. This can be seen in the following
extracts. In Task A, the relatively closed task, the students are
required to sort 20 vocabulary cards into semantic fields. In Task B,
having read a text on the topic of habits, the students are required to
have an open-ended discussion on the topic of bad habits. (Both
extracts are adapted from Nunan, 1991.)

Extract from Task A


Two students, Hilda and Carlos, are studying the following words
which have been typed onto pieces of cardboard. Their task is to
group the words together in a way which makes sense to them. There
is silence for several minutes as the students study the cards:
GEOGRAPHY, ASTRONOMY, AGRICULTURE, ECONOMICS, COMMERCE,
ENGLISH, SCIENCE, STATISTICS, BOOK, COMPUTER, PENCIL, DIARY,
NEWSPAPER, MAGAZINE, THAILAND, HONG KONG, MELBOURNE,
DARWIN, UNITED STATES, ASIAN, DIAGRAM, ILLUSTRATION, PICTURE,
CARTOON, VIDEO, COMPETENT, LAZY, INTERESTING, SUPERIOR,
UNCOMFORTABLE, REGION.
H: Statistic and diagram—they go together. You know diagram?
C: Yeah.
H: Diagram and statistic are family. . . but maybe, I think, statistic and
diagram—you think we can put in science? Or maybe. .
C: Science, astronomy, [yeah) and er can be agriculture.
H: Agriculture’s not a science.
C: Yes, it’s similar
H: No. . . . er may be Darwin and science….
C: What’s the Darwin?
H: Darwin is a man.
C: No, it’s one of place in Australia.
H: Yes, but it’s a man who discover something, yes, I’m sure.
C: OK.
H: And maybe, look, yes, picture, newspaper, magazine, cartoon,
book, illustration
C: [yeah). Maybe we can put lazy and English together. Er Hong Kong,
Thailand together. Asian. Er, United States. Diary with picture,
newspaper and so on. . . . Oh, I understand, look, look. Here, it’s only
adjective—lazy, competent, interesting and comfortable. Er, what is
it? Ah yes yes. (She begins to rearrange the cards.)
C: Darwin
Extract from Task B
Maria, Martha, Sylvia, and Sandy are taking part in a small-group
discussion on the topic of bad habits.
Maria: My next door neighbour . . . he make eh very noisy, very noisy
[yeah]. I can’t tell him because he’s very good people.
(The discussion continues for several minutes.)
Sylvia: . . . you don’t want to say anything because you might get
upset, of course. Me do the same thing because I’ve got neighbours in
my place and always you know do something I don’t like it but I don’t
like to say had because I think maybe, you, knowmake him upset
or….
Martha: I’ve got bad neighbour but I feel embarrass….
Sylvia: . . . to say something of course, like everyone….
Martha: They always come in and see what I’m doing—who’s coming.
[no good) [yeah, that’s no good) They want to check everything. If
they see I buy something from the market they expect me to give
them some. [oh yeah). [oh that’s not nice] But I ... it’s difficult.
Sylvia: It’s a difficult, yeah, but sometime it’s difficult ….
Martha: They can’t understand, I bought them and I gave money
(laughter) [yeah]
Martha: You know sometime difficult to the people because sometime
I can’t speak the proper, the language, and little bit hard to give to
understand . . . and that’s—sometime feel embarrass then, I can’t say
it, you know?
Maria: [turns to the fifth woman, who has not yet spoken) Sarah, you
tell [you tell now]
Sarah: My, er, for example, my sister in law she all the time snores in
her sleep [oh, yes) And my brother say, “Oh, I’m sorry, we must sleep
separate” [separate beds] (laughter). They did. [good idea) A good
idea because he couldn’t sleep. (Laughter.)
In addition to the fact that the different task types stimulated
different interactional patterns, the research also indicated that some
task types might be more appropriate than others for learners at
particular levels of proficiency. In the above study, it was found that
with lower-intermediate to intermediate learners, the relatively closed
tasks stimulate more modified interaction than relatively more open
tasks. This is not to say that such students should engage in closed
tasks to the exclusion of open tasks. The important thing is that
program planners and teachers should select a mix of tasks to reflect
the pedagogic goals of the curriculum.
Another element considered within task design is that of
teacher/learner roles. All pedagogic tasks contain roles for teachers
and learners, and conflict is likely to occur if there is a
misapprehension between teachers and learners about their
respective roles. Research related to learner roles has come up with
findings which run counter to the folk wisdom of the classroom. For
instance, Bruton and Samuda (1980) found that learners are capable
of correcting each other successfully. Additionally, according to Porter
(1986), learners produce more talk with other learners than with
native-speaking partners, and learners do not learn each other’s
errors. Finally, Gass and Varonis (1985) found that
there were advantages, when conducting groupwork, to pairing
learners of different proficiency levels as well as from different
language backgrounds.
The final element is that of setting, which refers to the learner
configuration (either teacher-fronted, small group, pair, or individual),
as well as the environment (whether the task takes place in the
classroom or outside the classroom). One of the first task studies to
be carried out, that by Long, Adams, and Castanos (1976), found that
small-group tasks prompt students to use a greater range of language
functions than teacher-fronted tasks.
In relation to environment, Montgomery and Eisenstein (1985) found
that supplementing classroom tasks with community-based
experiences resulted in significantly increased language gains.

FUTURE DIRECTIONS: EXTENDING THE RESEARCH AGENDA


Most of the research carried out during the eighties and described in
the preceding section was driven by Krashen’s input hypothesis,
which is based on the belief that opportunities for second language
acquisition are maximised when learners are exposed to language
which is just a little beyond their current level of competence
(Krashen, 1981, 1982). The central research issue here is, What
classroom tasks and patterns of interaction provide learners with the
greatest amount of comprehensible input? It has been argued that
patterns of interaction in which learners are forced to make
conversational adjustments
promote acquisition. As I have already pointed out, this view
represents an indirect rather than direct relationship between
environmental factors (for example, types of instruction) and
language acquisition. I also referred to research which, while
questioning the comprehensible input hypothesis, supported the
communicative tasks to which it gave rise.
While the research reviewed in the preceding section represents a
healthy state of affairs, the scope needs to be developed and
extended both substantively and methodologically. In substantive
terms, the research agenda needs to incorporate a greater range of
linguistic and psycholinguistic models. Methodologically, the scope of
the research needs to be extended by the utilization of a greater
range of research tools and techniques. In particular, it would be
useful to see the emergence of research which explored the
relationships between contextual factors, interpersonal factors,
learner proficiency levels; and pedagogic tasks.
In order to indicate the ways in which these principles might influence
the shape of future research, I shall briefly review two recent
investigations which provide useful indications of the ways in which
research on task-based language teaching and learning can be
extended both substantively and methodologically.
Berwick (1988; in press) exemplifies the advantages of extending the
research agenda on tasks by drawing on insights from a range of
theoretical models. Of particular interest is his utilization of functional
grammars, specifically the systemic-functional model first articulated
by Halliday (see, for example, Halliday, 1985; Halliday & Hasan, 1976;
Halliday & Hasan, 1989). This particular model of language attempts
to draw explicit links between the functions which language exists to
fulfil and its realization at the level
of lexicogrammatical choice. In his research, Berwick (1988) explored
differences at the level of lexicogrammar attributable to different task
types. In classifying tasks, he distinguished between pedagogical and
collaborative goals on one hand, and expository and experiential
processes on the other. Tasks with pedagogical goals are concerned
with the transfer of information through explicit instruction, while
collaborative tasks “emphasized cooperative, consensual behaviour
and exchange of information about a problem or topic which
participants explore freely during the task itself” (Berwick, in press).
Tasks based on expository processes are concerned with theoretically
based
knowledge, whereas experiential processes are concerned with
procedural knowledge (in familiar terms, the former is concerned with
theoretical knowing [knowing that], while the latter is concerned with
practical knowledge [knowing how]). Berwick uses these two
dimensions to situate a range of tasks which he used in his study.
These are set out in Figure 2. A description of the tasks follows.
Task COM1: This task, residing at the expository end of the process
continuum and the pedagogical end of the goal continuum, consisted
of a lecture about finding string characters in a text through use of
the word-processing program of a personal computer, not physically
present in the experimental setting.
Task COM2: This task shared the pedagogical goal of COM1, but was
more experiential in that it involved a demonstration of how to find
character strings on the laptop computer when it was physically
present in front of the participants.
Task LEC1: Participants in this task faced away from each other. One
participant had a small Lego toy made of snap-together plastic parts
which had to be described so that the second participant was able to
assemble a replica of the toy.
Task LEC2: This task was similar to LEG1, except that participants sat
face-to-face.
Task DIS: The Final task was an informal discussion of any topic of
common interest to the participants.
The independent variable in Berwick’s study was the task. Dependent
variables included a range of discourse features associated with the
negotiation of meaning in interaction and utilized in many of the task
investigations based on the input hypothesis. Variables included
clarification requests, comprehension checks, confirmation checks,
definitions, display questions, echoes, expressions of lexical
uncertainty, referential questions, selfexpansions,
self-repetitions, and other-repetitions.
FIGURE 2
Goal and Process Dimensions of the Five Tasks Used in the Berwick
Study
From “Towards an Educational Framework for Teacher-led Tasks in
English as a Foreign Language” by R.Berwick, in press, in Task-Based
Language Teaching edited by G. Crookes & S. Gass. Copyright
Multilingual Matters. Reprinted by permission.
Processes
Goals Expository Experiential
Pedagogical COM1 COM2
Collaborative/Soc LEG 1 DIS LEG2
ial
Berwick (in press) established through his research that task type is
an important determinant of lexicogrammatical exponents. He was
also able to relate the tasks and exponents to an educational
framework which provides a broad pedagogical rationale for task-
based language teaching. I have described his research at some
length because it exemplifies the value of drawing on a range of
theoretical models in the development of research programs into
task-based language teaching and learning.
The second study reviewed in this section is by Duff (in press). Duff
carried out a longitudinal case study of a single learner, investigating
the extent to which performance on different types of tasks yielded
different types of information on the subject’s interlanguage. The
three tasks investigated included an interview conversation, a picture
description, and a Cambodian folktale narration. The dependent
measures included the amount of language produced, the range of
vocabulary elicited, nominal reference, and negation. Data were
collected from a 24-year-old Cambodian male, over a 2-year period.
Duff’s study yielded mixed results. While there was some evidence of
task-related variability, the subject’s performance from one data-
collection period to the next also exhibited variability. The study
raised five fundamental questions: (a) Are the tasks selected distinct
enough to be operationalizable constructs in this type of analysis? (b)
Assuming the constructs are valid, are there any meaningful
differences across tasks? (c) To what extent can variability be
ascribed to other constructs such as genre or topic? (d) Were the
features investigated by the researcher the salient ones, or should
this line of research be restricted to those features of interlanguage
morphology and phonology
which have been found to be salient? (e) How is the researcher to
account for those differences which were observed?
Duff’s study is significant within the current context because it
represents a departure rom the cross-sectional research which has
typified the field since its inception. While the longitudinal case study
has been usefully employed in other aspects of SLA research (see, for
example, Schmidt, 1983) it is uncommon in research on tasks. In the
five fundamental questions she raises as a result of her study, Duff
also provides a basis for a substantial research agenda for further
research. Finally, she places the issue of interlanguage tt variability
firmly on the research agenda. Looking to the future, I would like to
see the issue of variability feature more prominently in research into
task-based language learning and teaching.

CONCLUSION
In this paper I have provided a selective overview of the development
of task-based language teaching. I have tried to show that, while it
had its genesis in mainstream education, task-based teaching has
become a powerful influence in language education. At a conceptual
level, the approach has been supported by changing conceptions of
the nature of language and learning—captured under the rubric of
communicative language teaching. Empirically, TBLT is supported by
a healthy research agenda which emerged from process-oriented
second language acquisition.
In the second part of the paper, I have tried to indicate some of the
directions that TBLT might take in the future. In particular, I have
suggested that the conceptual and empirical basis needs to be
extended both substantively and methodologically, and I described
two recent investigations which illustrate the possible shape of
research under such an extended agenda.

REFERENCES
Anderson, A., & Lynch, T. (1988). Listening. Oxford: Oxford University
Press.
Berwick, R. (1988). The effect of task variation in teacher-led groups
on repair of English as a foreign language. Unpublished doctoral
dissertation, University of British Columbia, Vancouver, Canada.
Berwick, R. (in press). Towards an educational framework for teacher-
led tasks in English as a foreign language. In C. Crookes & S. Gass
(Eds.), Task-based language teaching. Clevedon, Avon, England:
Multilingual Matters.
Breen, M. (1984). Process syllabuses for the language classroom. In C.
Brumfit (Ed.), General English syllabus design. Oxford: Pergamon
Press.
Breen, M. (1987). Learner contributions to task design. In C. Candlin &
D. Murphy (Eds.), Language learning tasks. Englewood Cliffs, NJ:
Prentice Hall.
Brindley, G. (1984). Needs analysis and objective setting in the Adult
Migrant Education Service. Sydney: Adult Migrant Immigration
Service.
Brown, C., & Yule, C. (1983). Teaching the spoken language.
Cambridge: Cambridge University Press.
Bruton, C., & Samuda, C. (1980). Learner and teacher roles in the
treatment of oral error in group work. PLELC Journal, 11(3), 49-63.
Doughty, C., & Pica, T. (1986). “Information gap” tasks: Do they
facilitate second language acquisition? TESOL Quarterly, 20(3), 305-
325.
Duff, P. (in press). Task force on interlanguage performance: An
analysis of task as independent variable. In C. Crookes & S. Cass
(Eds.), Task- based language teaching.
Clevedon, Avon, England: Multilingual Matters.
Gass, S., & Varonis, E. (1985). Task variation and nonnative/nonnative
negotiation of meaning. In S. Cass and C. Madden (Eds.), Input in
second language acquisition. Rowley, MA: Newbury House.
Halliday, M. A. K. (1985). An introduction to functional grammar.
London: Edward Arnold.
Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London:
Longman.
Halliday, M. A. K., & Hasan, R. (1989). Language, context and text:
Aspects of language in a social semiotic perspective. Oxford: Oxford
University Press.
Kohonen, V. (in press). Experiential language learning: Second
language learning as cooperative learner education. In D. Nunan
(Ed.), Collaborative language learning and teaching. Cambridge:
Cambridge University Press.
KoIb, D. (1984). Experiential learning: Experience as the source of
learning and
development. Englewood Cliffs, NJ: Prentice Hall.
Krashen, S. (1981). Second language acquisition and second language
learning. Oxford: Pergamon Press.
Krashen, S. (1982). Principles and practice in second language
acquisition. Oxford: Pergamon Press.
Larsen-Freeman, D., & Long, M. H. (in press). An introduction to
second language acquisition research. London: Longman.
Long, M. H. (1981). Input, interaction and second language
acquisition. In H. Winitz (Ed.), Native language and foreign language
acquisition (Annals of the New York Academy of Sciences No. 379, pp.
259-278). New York: New York Academy of Sciences.